SSA: simple scheduling algorithm for resilient packet ring networks

Abstract

The resilient packet ring (RPR), defined under IEEE 802.17, has been proposed as a high-speed backbone technology for metropolitan area networks. RPR is introduced to mitigate the underutilisation and unfairness problems associated with the current technologies SONET and Ethernet, respectively. The key performance objectives of RPR are to achieve high bandwidth utilisation, optimum spatial reuse on the dual rings, and fairness. The RPR standard implements three traffic classes: Class A, Class B, and Class C. The RPR MAC has one queue for each traffic class. A potential performance limitation is associated with the head-of-line blocking. When the MAC uses a single FIFO to buffer frames awaiting access, a packet that is traversing through a congestion point may block transmission of other packets destined to a point before the congestion. The use of virtual destination queues (VDQs) to avoid the head-of-line blocking is introduced. Different bandwidth allocation policies are discussed to assign rates to VDQs. Finally, a bandwidth allocation policy is proposed, which would achieve the maximum utilisation at a very low complexity.